[Re: NFS -current

[Dan Nelson]

In the last episode (Mar 25), Terry Lambert said:
> Ian Dowse wrote:
> > I'm not sure what you mean by a "lost" mount. Do all further
> > accesses to the filesystem hang?
> > 
> > It is normal enough to get the above 'not responding' errors
> > occasionally on a busy fileserver, but only if they are almost
> > immediately followed by 'is alive again' messages.
> 
> Particularly when using UDP with a "rsize" or "wsize" larger than the
> MTU, which Linux people do all the time.
> 
> As you are using UDP...
> 
> "If you hear hoofbeats, look for horses first, not zebras".

UDP works just fine on a switched network.  On my NFS servers I use an
8k rsize/wsize and UDP mounts on everything and have relatively few
dropped fragments.

Server 1 (NFS homedir server):
21:22  up 23 days,  8:02, 36 users, load averages: 0.00 0.01 0.00
ip:
        1783574152 total packets received
        520544166 fragments received
        82873817 packets reassembled ok 
        15246 fragments dropped after timeout

Server 2:
21:31  up 223 days, 13:17, 35 users, load averages: 0.19, 0.12, 0.09
ip:
        2089844841 total packets received (rolled over 4 times)
        3965873984 fragments received
        668066863 packets reassembled ok 
        25596 fragments dropped after timeout

(hmm whatever happened to the 64-bit network counters idea)

-- 
	Dan Nelson
	dnelson at allantgroup.com